A pioneering AI safety nonprofit that conducts research and public outreach to help prevent human extinction from the development of artificial superintelligence, with a current focus on policy advocacy and communications.
A pioneering AI safety nonprofit that conducts research and public outreach to help prevent human extinction from the development of artificial superintelligence, with a current focus on policy advocacy and communications.
People– no linked people
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- $7,100,000
- Current Runway
- 15 months
- Funding Goal
- $6,000,000
- Funding Raised to Date
- $48,000,000
Org Details
Updated 05/18/26The Machine Intelligence Research Institute (MIRI) is a 501(c)(3) nonprofit based in Berkeley, California, that conducts research and public outreach intended to help prevent human extinction from the development of artificial superintelligence (ASI). Founded in 2000 by Eliezer Yudkowsky, Brian Atkins, and Sabine Atkins as the Singularity Institute for Artificial Intelligence, the organization originally focused on accelerating AI development before pivoting in 2005 to focus on identifying and managing existential risks from superintelligent AI systems. In 2005 the institute moved from Atlanta to Silicon Valley and began to focus on ways to identify and manage risks from superintelligent AI, which were at the time largely ignored by scientists in the field. Starting in 2006, the institute organized the Singularity Summit to discuss the future of AI including its risks, initially in cooperation with Stanford University and with funding from Peter Thiel. MIRI's launch of LessWrong in 2009 under Yudkowsky's direction created a platform for discussing rationality techniques and AI risks. In January 2013, the institute rebranded as the Machine Intelligence Research Institute after selling its original name and the Singularity Summit to Singularity University. For most of its history, MIRI focused on technical alignment research, producing foundational work on topics including corrigibility, logical induction, and agent foundations. However, in January 2024, MIRI announced a major strategic pivot, concluding that alignment research had gone too slowly and was extremely unlikely to succeed in time to prevent catastrophe. The organization now pursues three objectives: policy work to secure international agreements halting progress toward ASI, communications to share MIRI's risk models with policymakers and the public, and continuing reduced-scale technical governance research. MIRI's current technical governance research explores questions bearing on regulatory and policy goals, including AI governance frameworks to avoid extinction, limitations of current AI evaluation methods, and verification mechanisms for international AI development agreements. In 2025, Eliezer Yudkowsky and Nate Soares published the book "If Anyone Builds It, Everyone Dies," arguing that the default consequence of creating artificial superintelligence is human extinction. MIRI's leadership includes Malo Bourgon as CEO, Nate Soares as President, Eliezer Yudkowsky as Co-Founder, Alex Vermeer as COO, and Jimmy Rintjema as CFO. Major funders have included Open Philanthropy (over $14.7 million since 2015), Vitalik Buterin, the Thiel Foundation, and Jaan Tallinn. The organization ran its first fundraiser in six years in December 2025, targeting $6 million.
Theory of Change
Updated 05/18/26MIRI believes that the default outcome of building artificial superintelligence is human extinction, and that technical alignment research alone is unlikely to succeed in time to prevent this. Their current theory of change centers on policy intervention: by communicating the extreme risks of ASI to policymakers, the public, and AI developers, MIRI aims to build support for a globally coordinated and collectively enforced moratorium on the development of ASI. In parallel, their technical governance research explores verification mechanisms for international agreements, limitations of current AI evaluation methods, and frameworks for AI governance that could reduce catastrophic risk. In the long term, MIRI hopes to see AI-empowered projects used to avert major AI mishaps while humanity develops the scientific and institutional maturity needed to make lasting decisions about the far future.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
Key risk: Their theory of change hinges on achieving an internationally coordinated ASI halt that is politically implausible in the near term, creating high execution and counterfactual impact risk and potential alienation of mainstream actors relative to more incremental governance or technical alignment approaches.
Case for funding: MIRI combines decades of credibility shaping alignment discourse with a singular focus on advocating a globally enforced ASI moratorium and producing verification/enforcement mechanisms, making them uniquely positioned to shift the Overton window and furnish the technical scaffolding for a hard pause if you share their 'doom by default' model.